Association Between Extracellular Volume Assessed by Cardiac MRI and New-Onset Atrial Fibrillation in Patients With ST-Segment Elevation Myocardial Infarction

Abstract

Objective

Although left ventricular (LV) fibrosis has been strongly linked to atrial fibrillation (AF), its relationship with new-onset AF (NOAF) following ST-segment elevation myocardial infarction (STEMI) remains unclear. This study aimed to investigate the association between different extracellular volume (ECV) measurements in the LV and NOAF during acute-phase STEMI.

Materials and Methods

This retrospective study included 517 patients diagnosed with acute STEMI (440 males, 77 females; mean age, 57.2 ± 12.4 years). All patients underwent cardiac magnetic resonance (CMR) imaging with T1 mapping sequences during hospitalization. Blood samples were collected within 24 hours of the CMR examination. ECV was assessed in three regions of the left ventricle: the non-myocardial infarction region (NMI-ECV), the myocardial infarction region (MI-ECV), and the entire myocardium (integral ECV). Multi-variable logistic regression was used to evaluate the associations between these ECV parameters and NOAF. Receiver operating characteristic (ROC) analysis was performed to assess the predictive value of ECV measurements, both alone and in combination with two conventional risk factors—N-terminal pro-B-type natriuretic peptide and infarct-related artery (right coronary artery).

Results

During hospitalization, 40 (7.7%) patients developed NOAF. After adjusting for confounding factors, including left atrial strain, MI-ECV, NMI-ECV, and integral ECV were independently associated with NOAF. The area under the ROC curve for predicting NOAF was 0.702 (95% confidence interval: 0.615–0.789), 0.625 (0.531–0.719), and 0.712 (0.627–0.798) for MI-ECV, NMI-ECV, and integral ECV, respectively. The addition of integral ECV and MI-ECV to conventional factors significantly improved the predictive performance for NOAF.

Conclusion

ECV measured using CMR was independently and significantly associated with NOAF occurrence in acute-phase STEMI. Incorporating ECV into risk assessment models could significantly improve NOAF prediction.

INTRODUCTION

Despite the widespread adoption of early reperfusion strategies, atrial fibrillation (AF) remains a common complication of acute myocardial infarction (AMI) [1]. The incidence of new-onset AF (NOAF) during hospitalization has been reported to range from 6.5% to 7.9% in patients undergoing thrombolysis or primary percutaneous coronary intervention (pPCI) [2,3]. Previous studies have demonstrated that any form of AF increases the risk of mortality, stroke, and heart failure in patients with AMI [4,5]. Additionally, the current guideline-recommended “triple antithrombotic therapy” presents challenges in managing AMI comorbid with AF due to the heightened risk of bleeding and poor adherence [6]. Therefore, identifying more relevant markers for predicting NOAF after AMI is essential.

One of the key mechanisms underlying AF development is left atrial (LA) structural remodeling, which can manifest as LA enlargement and may be a surrogate for left ventricular (LV) diastolic dysfunction [7]. In patients with myocardial infarction (MI), LV infarction can cause atrial stretch, hemodynamic changes, and neurohumoral activation, contributing to structural and functional atrial abnormalities [8]. Numerous studies have established a strong association between LV fibrosis and AF, with LV fibrosis also serving as a predictor of AF recurrence [9,10,11]. This suggests that myocardial fibrosis following AMI may serve as a valuable marker for NOAF. Extracellular volume (ECV), a noninvasive imaging biomarker, has been recognized as a surrogate for myocardial fibrosis [12]. Compared to late gadolinium enhancement (LGE) and T1 relaxation time, ECV provides a more reliable assessment of myocardial fibrosis during the acute phase of MI [13,14]. Previous research has indicated that increased ECV is associated with LA remodeling and serves as a strong independent predictor of AF recurrence [15]. However, the relationship between ECV and NOAF after AMI remains unclear. The primary objective of this study was to evaluate ECV in three LV regions, the non-myocardial infarction region (NMI-ECV), the MI region (MI-ECV), and the entire myocardium (integral ECV), using cardiac magnetic resonance (CMR) imaging. The study aimed to investigate the association between different LV ECV measurements and NOAF during the acute phase of ST-segment elevation myocardial infarction (STEMI).

MATERIALS AND METHODS

Study Population

This retrospective study included patients diagnosed with acute STEMI [1] between January 2020 and May 2024 (ClinicalTrials.gov ID. NCT06447857). All patients underwent CMR imaging during hospitalization, incorporating T1 mapping sequences. Blood samples were collected within 24 hours of the CMR examination. The inclusion criteria were: 1) Successful pPCI (TIMI flow ≥2) within 12 hours of symptom onset and 2) Admission to the cardiac care unit for continuous electrocardiogram (ECG) monitoring following pPCI. The exclusion criteria were: 1) Poor image quality, 2) History of previous MI, 3) History of AF, 4) Malignancy or inflammatory disease, 5) Severe valvular heart disease, and 6) Thyroid dysfunction. The study protocol was approved by the Institutional Review Board (IRB No. XYFY2024-KL189-01). Due to the minimal risk to patients, the requirement for written informed consent was waived in accordance with the relevant IRB guidelines. Clinical data, relevant laboratory indices, and medication details were obtained from patient records. NOAF was defined as post-admission AF without a prior history of the condition [16]. Continuous ECG monitoring revealed irregular rhythms during admission, which were confirmed as NOAF through a 12-lead ECG. Infarct-related artery (IRA) were identified based on coronary angiography findings. This study, involving 517 patients (440 males, 77 females; mean age, 57.2 ± 12.4 years), was larger and explored a different theme than our previously reported research [17]. The average length of hospitalization was 5.9 ± 1.9 days (Fig. 1).

Fig. 1. Study flowchart. STEMI = ST-segment elevation myocardial infarction, AF = atrial fibrillation.

CMR Imaging Protocol and Image Analysis

The median time to CMR completion was 5 (4–6) days after hospitalization. The CMR protocol and related parameters have been previously published [17,18]. Balanced turbo field echo sequence was adopted for imaging. Scan parameters included: slice thickness = 7 mm, no interlayer gap; echo time = 1.47 ms; repetition time = 2.94 ms; flip angle = 60°; field of view = 300 × 300 mm; matrix = 280 × 240 mm; and voxel size = 1.22 × 1.22 × 8.0 mm³. CMR imaging was performed using 3T systems (Ingenia, Philips, Eindhoven, The Netherlands). Image analysis was performed using CVI42 (cvi42^® version 5.13.5, Circle Cardiovascular Imaging, Calgary, Canada).

CMR image analysis was conducted independently by a radiologist (Z.L.) and a cardiologist (W.C.), each with 15 years of experience. Both were blinded to this study and performed repeat measurements, yielding four sets of data. The mean values were calculated and used for further analysis. LV parameters, including end-diastolic volume (EDV), end-systolic volume (ESV), and LV ejection fraction (LVEF), were automatically obtained and corrected for body surface area (BSA). The LA structure was analyzed by automatically tracing the LA border (except for the pulmonary veins and the left auricle). The biplane method was used to calculate atrial volume, LA volume = (0.85 × four-chamber LA area × two-chamber LA area)/shortest length of the LA vertical axis on two or four chambers with BSA correction. LA ejection fraction (LAEF) = ([Lamax - Lamin]/Lamax) × 100%. CMR imaging feature tracking was used to obtain the strain measurements, including reservoir strain, conduit strain, booster strain, and LV global longitudinal strain (GLS). All strain values were reported in absolute terms. The mean signal intensity and standard deviation (SD) of the myocardium in the region of interest (ROI) were measured, and regions with delayed enhancement were defined as areas where the signal intensity was ≥5 SD above the distal normal myocardial signal. Microvascular obstruction (MVO) was identified in the LGE images. The LGE area was measured, excluding MVO, and both LGE and MVO were expressed as percentages of total LV myocardial mass. Endocardial, epicardial, and blood pool regions were delineated on short-axis planes to assess the T1 relaxation time. Mean relaxation times (≥1 cm²) of two ROIs were plotted at the MI and NMI sites. Limbic and papillary muscles were carefully avoided, and the ROI was placed in the brightest infarcted area. Blood pool T1 value was derived from the LV blood pool. For the entire study population, routine ECV values were standardized for blood hematocrit (HCT) and calculated using the following equations: ECV = (1 - HCT) × (1/myocardial enhanced T1 - 1/myocardial native T1)/(1/blood pool enhanced T1 - 1/blood pool native T1) (Fig. 2).

Fig. 2. The measurement of ECV. A-D: Native T1 map (A), enhanced T1 map (B), generate ECV images (C), the distribution coefficient λ generates the image (D); red circle marks the endocardium, green circle marks the epicardium, ROI (yellow) marks the blood pool, ROI (pink) marks the non-myocardial infarction sites, ROI (blue) marks the myocardial infarction site. ECV = extracellular volume, ROI = region of interest.

Statistical Analysis

Statistical analysis was performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA) and R (Lucent Technologies, Murray Hill, NJ, USA). The Kolmogorov-Smirnov test was applied to assess data normality. Normally distributed continuous variables were expressed as mean ± SD and analyzed using Student's t-test. Continuous variables that were not normally distributed were expressed as median with interquartile range, and analyzed using the Mann-Whitney U test. Categorical variables were expressed as frequencies and percentages and analyzed using the chi-square test. Relationships between integral ECV and other CMR variables were assessed using Pearson correlation tests. Variables with a P-value <0.05 from univariable analyses were entered into the multi-variable logistic regression model using the stepwise forward method to identify factors associated with NOAF. Given the potential multicollinearity among the three ECV measurements (integral ECV, MI-ECV, and NMI-ECV), each was analyzed separately in three distinct multi-variable logistic regression models. Restricted cubic splines were used to explore the dose-response relationship between ECV and NOAF. A conventional logistic regression model incorporating N-terminal pro-B-type natriuretic peptide (NT-proBNP) and IRA-right coronary artery (RCA) was developed for predicting NOAF based on multi-variable analysis and previous studies [8,17,19]. Additional combined logistic regression models were created by incorporating integral ECV, MI-ECV, and NMI-ECV into the conventional model. Receiver operating characteristic (ROC) was used to assess the performance of ECV measurements and predictive models in identifying NOAF. The DeLong test was used to compare the areas under ROC curves (AUCs). The net reclassification index (NRI) and integrated discrimination improvement (IDI) were also used to evaluate improvements in model discrimination and patient reclassification. The intraclass correlation coefficient was used to assess measurement consistency both within a single observer and between different observers (inter- and intra-observer). All P-values were from two-sided tests, and statistical significance was set at P < 0.05.

RESULTS

Baseline Patient Characteristics and CMR Results

A total of 517 patients with STEMI were enrolled in this study, including 40 (7.7%) patients with NOAF. The inter- and intra-observer agreement for each ECV measurement is detailed in Supplementary Table 1. Compared to the non-AF group, patients with NOAF exhibited higher levels of high-sensitivity troponin T (hs-TnT), NT-proBNP, LV-ESV index (LV-ESVi), LA-EDV index (LA-EDVi), LGE%, MVO%, and proportion of patients with RCA and lower values of systolic blood pressure (SBP), LVEF, LAEF, GLS, and LA strain. Additionally, the MI-ECV (51.0% ± 10.7% vs. 43.2% ± 10.5%, P < 0.001), NMI-ECV (27.1% ± 6.2% vs. 24.1% ± 5.3%, P < 0.001), and integral ECV (33.6% ± 5.9% vs. 29.0% ± 5.9%, P < 0.001) were significantly higher in the AF group than in the non-AF group (Tables 1, 2).

Table 1. Patient characteristics.

Characteristic		Total (n = 517)	Non-AF (n = 477)	AF (n = 40)	P
Age, yrs		57.2 ± 12.4	57.0 ± 12.3	60.2 ± 13.7	0.119
Sex, female		77 (14.9)	71 (14.9)	6 (15.0)	0.984
BMI, kg/m2		25.7 ± 3.3	25.7 ± 3.3	25.2 ± 2.6	0.337
SBP, mmHg		128.1 ± 18.7	128.6 ± 19.0	122.5 ± 14.8	0.048
DBP, mmHg		80.3 ± 13.4	80.3 ± 13.3	79.2 ± 13.9	0.595
Heart rate, bpm		79.1 ± 13.1	78.9 ± 13.1	81.3 ± 13.6	0.258
Smoking		275 (53.2)	258 (54.1)	17 (42.5)	0.158
Hypertension		231 (44.7)	216 (45.3)	15 (37.5)	0.342
Diabetes mellitus		128 (24.8)	120 (25.2)	8 (20.0)	0.468
Stroke		61 (11.8)	57 (12.0)	4 (10.0)	0.911
IRA
	LCX	104 (20.1)	98 (20.6)	6 (15.0)	0.401
	LAD	256 (49.5)	243 (50.9)	13 (32.5)	0.025
	RCA	153 (29.6)	133 (27.9)	20 (50.0)	0.003
	Others	4 (0.8)	3 (0.6)	1 (2.5)	0.276
Killip class					0.862
	I	476 (92.1)	439 (92.0)	37 (92.5)
	II	17 (3.3)	15 (3.1)	2 (5.0)
	III	6 (1.2)	6 (1.3)	0 (0.0)
	IV	18 (3.5)	17 (3.6)	1 (2.5)
Medication
	Statins	492 (95.2)	456 (95.6)	36 (90.0)	0.230
	Dapagliflozin	140 (27.1)	130 (27.3)	10 (25.0)	0.758
	ACEI/ARB	323 (62.5)	301 (63.1)	22 (55.0)	0.309
	β-blockers	437 (84.5)	401 (84.1)	36 (90.0)	0.319
	Spironolactone	47 (9.1)	41 (8.6)	6 (15.0)	0.286
Laboratory indicators
	Total cholesterol, mmol/L	4.34 ± 1.03	4.34 ± 1.04	4.30 ± 0.84	0.814
	Triglycerides, mmol/L	1.77 ± 1.22	1.77 ± 1.21	1.84 ± 1.33	0.711
	HDL-C, mmol/L	0.94 ± 0.23	0.94 ± 0.23	0.95 ± 0.23	0.888
	LDL-C, mmol/L	2.76 ± 0.87	2.76 ± 0.88	2.68 ± 0.82	0.583
	FBG, mmol/L	6.51 ± 2.56	6.50 ± 2.59	6.69 ± 2.09	0.648
	eGFR, mL/min/1.73 m2	108.9 ± 15.2	109.1 ± 14.9	106.1 ± 17.5	0.230
	Peak hs-TnT, ng/L	2478.0 (886.7, 5319.0)	2345.0 (866.0, 5118.0)	4676.5 (1892.5, 8159.8)	0.011
	Peak NT-proBNP, pg/mL	1055.0 (510.0, 2143.2)	998.0 (495.0, 1986.0)	2174.0 (1262.5, 3190.7)	<0.001
	Peak hs-CRP, mg/L	13.9 (4.4, 35.9)	13.8 (4.3, 35.1)	16.1 (7.6, 44.8)	0.398

Data are presented as count (%) for categorical variables and median (Q1, Q3) or mean ± standard deviation for continuous variables.

AF = atrial fibrillation, BMI = body mass index, SBP = systolic blood pressure, DBP = diastolic blood pressure, IRA = infarct-related artery, LCX = left circumflex artery, LAD = left atrial diameter, RCA = right coronary artery, Others = left main coronary artery and intermediate branch, ACEI = angiotensin-converting-enzyme inhibitor, ARB = angiotensin II receptor blocker, HDL-C = high-density leptin cholesterol, LDL-C = low-density leptin cholesterol, FBG = fasting blood glucose, eGFR = estimated glomerular filtration rate, hs-TnT = high sensitivity troponin T, NT-proBNP = N-terminal pro-B-type natriuretic peptide, hs-CRP = high sensitivity C-reactive protein

Table 2. CMR results.

CMR parameter	Total (n = 517)	Non-AF (n = 477)	AF (n = 40)	P
MI-ECV, %	43.8 ± 10.7	43.2 ± 10.5	51.0 ± 10.7	<0.001
NMI-ECV, %	24.3 ± 5.4	24.1 ± 5.3	27.1 ± 6.2	<0.001
Integral ECV, %	29.3 ± 6.1	29.0 ± 5.9	33.6 ± 5.9	<0.001
LVEF, %	46.8 ± 10.4	47.3 ± 10.2	40.9 ± 10.5	<0.001
LAEF, %	48.9 ± 13.0	49.4 ± 12.9	42.4 ± 13.0	0.001
GLS, %	13.5 ± 4.3	13.7 ± 4.2	10.7 ± 4.2	<0.001
LV-EDVi, mL/m2	78.2 ± 19.8	77.7 ± 19.1	83.5 ± 26.2	0.181
LV-ESVi, mL/m2	43.7 ± 16.6	43.0 ± 15.9	52.1 ± 22.1	0.014
LV mass, g	111.5 ± 24.3	111.1 ± 23.8	116.7 ± 29.9	0.163
LGE%	25.0 ± 12.8	24.5 ± 12.6	31.0 ± 13.9	0.002
MVO%	0 (0, 1.2)	0 (0, 1.2)	0.2 (0, 3.8)	0.048
LA reservoir strain, %	24.0 ± 9.4	24.8 ± 9.1	14.6 ± 8.0	<0.001
LA booster strain, %	11.1 ± 5.7	11.4 ± 5.7	6.9 ± 4.8	<0.001
LA conduit strain, %	12.9 ± 6.3	13.3 ± 6.2	7.8 ± 4.6	<0.001
LA-EDVi, mL/m2	18.4 ± 8.8	18.1 ± 8.5	21.6 ± 10.7	0.048
LA-ESVi, mL/m2	35.7 ± 13.8	35.7 ± 14.0	35.0 ± 12.2	0.758

Data are presented as median (Q1, Q3) or mean ± standard deviation for continuous variables.

CMR = cardiac magnetic resonance, AF = atrial fibrillation, MI = myocardial infarction, ECV = extracellular volume, NMI = non-myocardial infarction, LVEF = left ventricular ejection fraction, LAEF = left atrial ejection fraction, GLS = left ventricular global longitudinal strain, LV = left ventricular, EDVi = end-diastolic volume index, ESVi = end-systolic volume index, LGE = late gadolinium enhanced, MVO = microvascular obstruction, LA = left atrial

Relationship Between ECV and NOAF

Uni-variable logistic regression analysis identified MI-ECV, NMI-ECV, integral ECV, GLS, LVEF, LV-ESVi, LGE%, MVO%, LA reservoir strain, LA booster strain, LA conduit strain, LAEF, LA-EDVi, SBP, hs-TnT, NT-proBNP, and IRA-RCA as factors associated with NOAF. Three multi-variable logistic regression models (models 1, 2, and 3 incorporating MI-ECV, NMI-ECV, and integral ECV, respectively) were constructed. After adjusting for confounding factors, NT-proBNP, IRA-RCA, LA reservoir strain, MI-ECV (odds ratio [OR] = 1.09; 95% confidence interval [CI]: 1.05–1.13, P < 0.001), NMI-ECV (OR = 1.07; 95% CI: 1.00–1.14, P = 0.036), and integral ECV (OR = 1.11 95% CI: 1.05–1.19, P = 0.001) remained significantly correlated with NOAF (Table 3, Supplementary Table 2). A linear dose-response relationship was observed between ECV and NOAF (P for nonlinear >0.05), suggesting that the risk of NOAF increases with elevated levels of ECV (Fig. 3).

Table 3. Association with NOAF: multi-variable logistic regression analysis.

Variables	MI-ECV (model 1)		NMI-ECV (model 2)		Integral ECV (model 3)
	OR (95% CI)	P	OR (95% CI)	P	OR (95% CI)	P
MI-ECV	1.09 (1.05–1.13)	<0.001	Not in model		Not in model
NMI-ECV	Not in model		1.07 (1.00–1.14)	0.036	Not in model
Integral ECV	Not in model		Not in model		1.11 (1.05–1.19)	0.001
LA reservoir strain	0.85 (0.80–0.90)	<0.001	0.86 (0.81–0.91)	<0.001	0.86 (0.81–0.91)	<0.001
Peak NT-proBNP	1.48 (1.13–1.93)	0.004	1.36 (1.06–1.74)	0.018	1.35 (1.05–1.75)	0.021
IRA-RCA	3.04 (1.42–6.51)	0.004	2.69 (1.29–5.59)	0.008	2.81 (1.33–5.92)	0.007

Model 1 initially included all variables with P < 0.05 in uni-variable logistic regression analysis except NMI-ECV and Integral ECV. Model 2 initially included all variables with P < 0.05 in uni-variable logistic regression analysis except MI-ECV and Integral ECV. Model 3 initially included all variables with P < 0.05 in uni-variable logistic regression analysis except MI-ECV and NMI-ECV. The variables in the final models are only shown in the table, which specify the logistic equations used to compute the AUC.

NOAF = new-onset atrial fibrillation, NMI = non-myocardial infarction, ECV = extracellular volume, MI = myocardial infarction, AUC = area under the curve, OR = odds ratio, CI = confidence interval, LA = left atrial, NT-proBNP = N-terminal pro-B-type natriuretic peptide, IRA = infarct-related artery, RCA = right coronary artery

Fig. 3. Dose-response relationship between ECV and NOAF in patients with STEMI. A-C: A dose-response relationship between integral ECV and NOAF (A), a dose-response relationship between MI-ECV and NOAF (B), a dose-response relationship between NMI-ECV and NOAF (C). ECV = extracellular volume, NOAF = new-onset atrial fibrillation, STEMI = ST-segment elevation myocardial infarction, MI = myocardial infarction, NMI = non-myocardial infarction, CI = confidence interval.

ECV for Identifying NOAF

ROC analysis was performed to evaluate the predictive performance of NMI-ECV, MI-ECV, and integral ECV for NOAF. The AUC values for MI-ECV, NMI-ECV, and integral ECV were 0.702 (95% CI: 0.615–0.789), 0.625 (0.531–0.719), and 0.712 (0.627–0.798), respectively (Table 4).

Table 4. Performance of ECV measurements for predicting new-onset atrial fibrillation.

ECV measurement	AUC	95% CI	Cut-off	Sensitivity	Specificity
MI-ECV	0.702	0.615–0.789	52.6	0.550 (22/40)	0.818 (390/477)
NMI-ECV	0.625	0.531–0.719	26.8	0.475 (19/40)	0.746 (356/477)
Integral ECV	0.712	0.627–0.798	32.3	0.625 (25/40)	0.746 (356/477)

ECV = extracellular volume, AUC = area under the curve, CI = confidence interval, MI = myocardial infarction, NMI = non-myocardial infarction

The AUC of the conventional model (which included NT-proBNP and IRA-RCA) for predicting NOAF was 0.725 (0.651–0.799). Adding integral ECV to the conventional model increased the AUC to 0.786 (P < 0.029 compared to the conventional model). Adding MI-ECV to the conventional model increased the AUC to 0.792 (P < 0.046 compared to the conventional model). Adding NMI-ECV to the conventional model increased the AUC to 0.749 (P = 0.261 compared to the conventional model). MI-ECV and integral ECV consistently significantly improved NOAF prediction compared to the conventional model, as evaluated by AUC, NRI, and IDI, while NMI-ECV did not (Fig. 4, Table 5).

Fig. 4. Receiver operating characteristic analysis of combined parameters for identifying new-onset atrial fibrillation. IRA = infarct-related artery, RCA = right coronary artery, NT-proBNP = N-terminal pro-B-type natriuretic peptide, AUC = area under the curve, MI = myocardial infarction, ECV = extracellular volume, NMI = non-myocardial infarction.

Table 5. ECV combined with conventional parameters for identifying new-onset atrial fibrillation.

Models*	ROC		NRI		IDI
	AUC (95% CI)	P †	Estimate (95% CI)	P †	Estimate (95% CI)	P †
Conventional model	0.725 (0.651–0.799)	-	Reference	-	Reference	-
Conventional model + MI-ECV	0.792 (0.719–0.865)	0.046	0.537 (0.229–0.845)	<0.001	0.045 (0.023–0.067)	<0.001
Conventional model + NMI-ECV	0.749 (0.672–0.825)	0.261	0.149 (-0.174–0.471)	0.365	0.021 (0.004–0.038)	0.018
Conventional model + integral ECV	0.786 (0.716–0.856)	0.029	0.537 (0.225–0.849)	<0.001	0.047 (0.021–0.072)	<0.001

^*The models are logistic regression models, and the conventional model included right coronary artery and N-terminal pro-B-type natriuretic peptide, ^†Comparison with the conventional model.

ECV = extracellular volume, ROC = receiver operating characteristic, NRI = net reclassification index, IDI = integrated discrimination improvement, AUC = area under the curve, CI = confidence interval, MI = myocardial infarction, NMI = non-myocardial infarction

DISCUSSION

To the best of our knowledge, this study is the first to investigate the relationship between ECV measurements in different LV regions and NOAF in patients with STEMI. The main findings of this study are as follows: 1. Elevated ECV was an independent risk factor for NOAF, even after adjusting for potential confounders, including LA strain. 2. A dose-response relationship was observed, indicating that NOAF risk increases with higher ECV levels. 3. Adding ECV measurements to conventional parameters significantly improved the predictive performance for NOAF.

Patients with any form of AF in AMI face an increased risk of adverse cardiovascular events, including those who develop NOAF during hospitalization [4,5]. In our study, the incidence of NOAF during hospitalization was 7.7%, which may be influenced by various factors. Previous research has identified inflammation, myocardial fibrosis, atrial ischemia, and LV dysfunction as crucial mechanisms contributing to AF development post-MI [20,21,22].

Myocardial fibrosis is a major contributor to poor prognosis in patients with MI [23]. In addition to its role in atrial remodeling, previous studies have established a causal relationship between AF and LV fibrosis [9,10,11]. With the advancements in imaging modalities, CMR imaging has become a crucial tool for assessing cardiac structure and function in patients with AMI [24,25,26]. Ling et al. [27] demonstrated an association between AF and diffuse LV myocardial fibrosis by analyzing LV T1 relaxation time. In another study, Ling & Tandri [28] found that 13% of patients with AF without a history of MI exhibited LGE on CMR, with higher LGE levels correlating with increased mortality risk. ECV is defined as the coefficient of change in tissue and blood T1 values before and after contrast injection, serving as a valuable marker for myocardial fibrosis. Compared to T1 relaxation time and LGE, ECV offers a more reliable assessment of myocardial fibrosis by minimizing the influence of intracellular pathology [29,30]. Although ECV is widely used for risk stratification in patients with AMI as a surrogate for myocardial fibrosis, studies exploring the relationship between ECV and AF remain limited. A previous study involving patients with hypertension and AF undergoing pulmonary vein isolation found that elevated ECV was associated with diastolic dysfunction and LA remodeling, making it a strong independent predictor of recurrent AF [15]. However, the link between ECV and NOAF after AMI remains unclear. This study is the first to identify elevated ECV as an independent risk factor for NOAF after STEMI, demonstrating a dose-response relationship between ECV levels and NOAF risk. Our findings provide strong evidence supporting the association between LV fibrosis and NOAF in patients with STEMI. A potential explanation for this association is that higher ECV reflects reduced LV function, leading to elevated LV end-diastolic pressure, ventricular stiffness, and diastolic dysfunction, which in turn increases atrial load and potentially triggers AF [31]. Additionally, evidence suggests that LV infarction can induce atrial stretch, hemodynamic alterations, and neurohumoral activation, contributing to atrial abnormalities following MI [8]. Therefore, the association between ECV and NOAF may be mediated through ventricular-atrial interactions. Supporting this hypothesis, our study found a correlation between integral ECV and several CMR ventricular parameters (Supplementary Fig. 1). GLS is a well-recognized sensitive and reliable marker of cardiac function in the early stages of AMI [32], and in our study, it was significantly associated with NOAF, partially explaining our findings. Moreover, integral ECV was also found to be associated with some key CMR atrial parameters (Supplementary Fig. 1). Furthermore, a previous study identified reduced LA strain as a strong risk factor for NOAF after STEMI [17], suggesting that the observed correlation between integral ECV and LA strain may offer additional insights to our findings.

NMI-ECV has recently become a research hotspot, with several studies highlighting its association with prognosis following MI [14,33,34]. Consistent with these studies, our study demonstrated a significant association between NMI-ECV and NOAF during hospitalization. This association may be explained by the acute stress response triggered in the early stages of STEMI, which promotes local cytokine production from cardiomyocytes and macrophages, leading to maladaptive matrix modifications and collagen deposition associated with impaired distal myocardial contractility [35,36]. A previous study observed a marked increase in NMI-ECV in patients with STEMI treated with pPCI compared to healthy controls [33]. Additionally, in a mouse model of AMI, Tsuda et al. [37] found molecular and immunohistochemical evidence of distal interstitial fibrosis as early as 72 hours post-infarction. Interestingly, an earlier STEMI study [21] reported no significant differences in myocardial salvage, infarct size, or microvascular injury between patients with and without AF. The discrepancy may stem from differences in AF definitions: AF was defined in our study as a newly developed event post-STEMI, whereas their study included patients with pre-existing AF. Moreover, our study highlighted the relevance of ECV as a valuable imaging biomarker. Notably, acute ECV has been shown to provide a more accurate assessment of infarct size than acute LGE [38]. These findings suggest that NMI-ECV may hold potential as a novel prognostic marker for STEMI. However, further research is needed to validate these findings.

Consistent with previous research, our study identified LA strain, IRA-RCA, and NT-proBNP as independent risk factors for NOAF [8,17,19]. Notably, incorporating either integral ECV or MI-ECV into the conventional model significantly improved its predictive performance for NOAF. Our study provides evidence supporting the role of ventricular fibrosis in NOAF development. Since early myocardial fibrosis in patients with AMI can be mitigated with appropriate treatment, our study may offer a new direction for preventing and treating NOAF post-AMI. Moreover, combining ECV with conventional clinical indicators could enhance risk stratification for patients with STEMI regarding NOAF.

This study has some limitations. First, as a retrospective study, it is subject to potential unavoidable bias. Second, our study could not establish a strict causal relationship between ECV and NOAF, which would require a prospective design and a larger sample size. However, the primary aim of this study was to explore the association between ECV and NOAF. Third, the study was limited to patients with STEMI, so the findings may not be directly generalizable to other populations. Fourth, the small number of NOAF events makes our results exploratory, necessitating validation using larger cohorts. Fifth, our study focused only on the acute phase of STEMI and NOAF during hospitalization. Due to the absence of long-term follow-up, the predictive value of ECV for NOAF after hospital discharge remains unclear. Lastly, the comparisons between models are preliminary, as our study lacks both internal and external validations.

In conclusion, ECV measured with CMR was identified as an independent risk factor for NOAF in patients with acute-phase STEMI. ECV could significantly improve NOAF prediction, and its assessment using CMR may offer valuable insights into NOAF occurrence after STEMI.

Availability of Data and Material

The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.

Supplement

The Supplement is available with this article at https://doi.org/10.3348/kjr.2025.0070.